Google Plans to Directly Sell TPUs to Meta, NVIDIA's Market May Face Further Erosion
BlockBeats News, November 25th, according to foreign media reports, Google's parent company Alphabet (GOOG.O) is in talks with Meta (META.O) and other companies, planning to open up the usage rights of its self-developed Tensor AI chip to them, further expanding the competitive landscape with NVIDIA (NVDA.O). In response to this news, Google and its AI chip partner Broadcom rose after hours, while NVIDIA and AMD's stock prices fell accordingly. Google has traditionally only deployed its self-developed TPUs in its own data centers and rented out computing power to customers.
However, the U.S. tech media The Information reported on Monday night local time that Google is now planning to directly sell TPU chips to customers for deployment in their own data centers. Meta is considering purchasing Google's TPU chips worth billions of dollars starting in 2027 for use in its data centers, while also planning to start renting TPU computing power from Google Cloud as early as 2026. Currently, Meta's AI business mainly relies on NVIDIA GPUs. For Google and Broadcom, who jointly designed the Tensor chip, this signifies a potentially huge emerging market. It may also pose significant competitive pressure on NVIDIA and AMD, potentially impacting the latter's sales and pricing power. (Jinse)
You may also like

Who Will Control AI? Why Decentralized AI May Be the Only Alternative to Government and Big Tech
AI has become critical infrastructure, and governments and corporations are competing to control it. Centralized development and regulation are entrenching existing power structures. The Web3 community is building a decentralized alternative — distributed compute, token incentives, and community governance — before that window closes.

Vitalik wrote a proposal teaching you how to secretly use AI large models

On the eve of the explosion of on-chain options

WEEX AI Hackathon: How Did This AI Trading Winner Succeed?
A self-taught AI trading enthusiast achieved top-10 results at the WEEX AI Hackathon. Learn about the mindset, AI tools, and lessons behind this impressive performance.

One Balance to Rule Them All: Gravitas' On-Chain Prime Broker Ambition

That person who cashed out at the NFT peak is now selling a new shovel in the OpenClaw craze

Inter-generational Prisoner's Dilemma Resolution: The Nomadic Capital and Bitcoin's Inevitable Path

Upstream and downstream are starting to fight, all for the sake of everyone being able to "Lobster"

Circle and Mastercard Announce Partnership, the Next Stage for the Crypto Industry Belongs to Payments

From 5 Mao per kWh of Chinese electricity to a $45 API export: Tokens are rewriting currency units

Why is OpenAI playing catch-up to Claude Code instead?

Vitalik wrote a proposal teaching you how to secretly use AI large models

The doubling of Circle's stock price and the paradigm shift of stablecoins

Key Market Information Discrepancy on March 13th - A Must-See! | Alpha Morning Report

On-Chain Options Explosion.ActionEvent

《Time》 Magazine Names Anthropic as the World's Most Disruptive Company

Predictions market gains mainstream traction in the US, Canada, Claude launches Chart Interaction feature, What's the English community talking about today?

500 Million Dollars, 12 Seconds to Zero: How an Aave Transaction Fed Ethereum's "Dark Forest" Food Chain
Who Will Control AI? Why Decentralized AI May Be the Only Alternative to Government and Big Tech
AI has become critical infrastructure, and governments and corporations are competing to control it. Centralized development and regulation are entrenching existing power structures. The Web3 community is building a decentralized alternative — distributed compute, token incentives, and community governance — before that window closes.
Vitalik wrote a proposal teaching you how to secretly use AI large models
On the eve of the explosion of on-chain options
WEEX AI Hackathon: How Did This AI Trading Winner Succeed?
A self-taught AI trading enthusiast achieved top-10 results at the WEEX AI Hackathon. Learn about the mindset, AI tools, and lessons behind this impressive performance.